feat(bigframes): Support Expression objects in create_model options

This change allows the `options` parameter of `bigframes.bigquery._operations.ml.create_model` to accept BigFrames `Expression` objects. These expressions are compiled to SQL scalar expressions and included in the generated `CREATE MODEL` DDL statement.

- Added `bigframes.core.expression.Expression` type support in the `options` dict.
- Updated `create_model_ddl` to handle compiling expressions using `expression_compiler`.
- Added `test_create_model_expression_option` snapshot test to verify the generated "golden SQL", using an expression that calls a function on a literal value (e.g. 0.1 * 10).

Co-authored-by: tswast <247555+tswast@users.noreply.github.com>

Author: google-labs-jules[bot]

packages/bigframes/tests/unit/core/sql/snapshots/test_ml/test_create_model_expression_option/create_model_expression_option.sql

@@ -1,3 +1,3 @@
 CREATE MODEL `my_model`
-OPTIONS(l2_reg = 0.1, booster_type = 'gbtree')
+OPTIONS(l2_reg = 0.1 * 10, booster_type = 'gbtree')
 AS SELECT * FROM t

packages/bigframes/tests/unit/core/sql/test_ml.py

@@ -99,11 +99,19 @@ def test_create_model_list_option(snapshot):
 
 def test_create_model_expression_option(snapshot):
     import bigframes.core.expression as ex
+    import bigframes.operations.numeric_ops as numeric_ops
+    import bigframes.dtypes as dtypes
+
+    # An expression that calls a function on a literal value
+    # e.g. 0.1 * 10
+    literal_expr = ex.ScalarConstantExpression(0.1, dtypes.FLOAT_DTYPE)
+    multiplier_expr = ex.ScalarConstantExpression(10, dtypes.INT_DTYPE)
+    math_expr = ex.OpExpression(op=numeric_ops.mul_op, inputs=(literal_expr, multiplier_expr))
 
     sql = bigframes.core.sql.ml.create_model_ddl(
         model_name="my_model",
         options={
-            "l2_reg": ex.ScalarConstantExpression(0.1, None),
+            "l2_reg": math_expr,
             "booster_type": "gbtree",
         },
         training_data="SELECT * FROM t",